Hi Jon,
Cluster is capable of fantastic performance, and its also capable of
disappointing performance. If you've worked with it much, you might
already know this. Sometimes people expect tuning and administration
to be as simple as MySQL, and they aren't.
Hundreds of concurrent operations per second at 90% reads should be
easy to acheive, but the hard part might be getting the latency down
to an acceptable level.
My own advice is like this: First, use 5.1. There's not really much
reason to use 5.0 Cluster anymore. Second, Dolphin SCI cards cost as
little as $1500 per box and can be well worth it.
In a web environment, the latency (page load time) is often an
important performance metric. And MySQL/InnoDB, which fetches data
from a local disk on the database server, can turn it around much
faster than MySQL Cluster, where the data is over the network on an
NDB node. This is the problem that prompted me to start my open
source project mod_ndb, which you also might want to look at --
http://mod-ndb.googlecode.com/
JD
On Nov 13, 2007, at 9:39 PM, Jonathan Haddad wrote:
> Has anyone on the list set up cluster for use in a web environment?
> How
> does the performance compare to innodb for selects based off an index?
>
> How well does cluster perform with a lot of mixed reads / writes?
> (say 20
> to 1?) If a table is constantly having rows inserted, are there any
> locking
> issues we should be aware of? Usually we deal with this by
> partitioning
> our tables into multiple machines. Is this necessary anymore?
>
> Initially, there won't be much on here, but I'm expecting it to be
> running
> this on 6x 16GB dual core boxes for the data nodes - will 1 table
> with 500
> million rows and several hundred concurrent selects / inserts per
> second be
> an issue?
>
> Thanks,
> Jon
> --
> Jon Haddad
> jon@stripped
> Letsgetnuts.com
